Presentation is loading. Please wait.

Presentation is loading. Please wait.

Scheduling in IEEE 802.16e Mobile WiMAX Networks-Key Issues and a Survey 報告者 : 李宗穎 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 27, NO. 2, FEBRUARY.

Similar presentations


Presentation on theme: "Scheduling in IEEE 802.16e Mobile WiMAX Networks-Key Issues and a Survey 報告者 : 李宗穎 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 27, NO. 2, FEBRUARY."— Presentation transcript:

1 Scheduling in IEEE 802.16e Mobile WiMAX Networks-Key Issues and a Survey 報告者 : 李宗穎 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 27, NO. 2, FEBRUARY 2009 Chakchai So-In, Student Member, IEEE, Raj Jain, Fellow, IEEE, and Abdel-Karim Tamimi, Student Member, IEEE

2 2 Outline Introduction Scheduler Classification of Scheduler Conclusion

3 3 Introduction Key Features of WiMAX Networks WiMAX QoS Service Classes Application Traffic Models Request/Grant Mechanism

4 4 Key Features of WiMAX Networks OFDMA Scalable use of any spectrum width TDD and FDD MIMO Per subscriber adaptive modulation Advanced coding techniques (space-time coding and turbo coding) Strong security Multiple QoS classes

5 5 WiMAX QoS Service Classes Unsolicited Grant Scheme (UGS) Extended Real Time Polling Service (ertPS) Real Time Polling Service (rtPS) Non Real Time Polling Service (nrtPS) Best Effort Service (BE)

6 6 Application Traffic Models

7 7 Online-Game Traffic Model Quake II Traffic Model CDF : F(x) = 1 – e -e (x-a)/b inverse transformation : x = a + blnlnu

8 8 Detail Parameter

9 9 Request/Grant Mechanism (UGS) the scheduler needs to be aware of the resource requirements and should be able to schedule the flows so that the resources can be optimized EX : each flow requiring 500 bytes every 5 frames, if only 2500 bytes are allowed in one frame, all 10 flows can not start in the same frame, the problem gets more difficult when the UGS flows dynamically join and leave

10 10 Request/Grant Mechanism (rtPS) There is a strict or loose requirement of delay. If any packets are over the deadline, those packets will be dropped Video applications also have their own characteristics such as the size and the duration of I-frame, B-frame and P-frame for MPEG video

11 11 Request/Grant Mechanism (ertPS) Adaptive Multi-Rate (AMR) coding is used, only 33 bytes are sent every 20 ms during the active periods and 7 bytes during silent periods how to let the BS know whether there is a packet to transmit or not reduce the polling overhead and delay

12 12 Request/Grant Mechanism (nrtPS & BE) The only constraint for nrtPS is the minimum guaranteed throughput The main issue for BE is fairness. The problem is whether the scheduler should be fair in a short-term or a longterm EX : over one second, a flow can transmit 1 byte every 5 ms or 200 bytes every 1 second. Also, the scheduler should prevent starvation

13 13 Component Schedulers at BS and MSs

14 14 QoS Parameters The main parameters are the minimum reserved traffic, the maximum allowable delay and the tolerated jitters the scheduler may need to reschedule or interleave packets in order to meet the delay and throughput requirements. (Ex : Earliest Deadline First)

15 15 Throughput Optimization The bandwidth request is indicated in number of bytes and it does not translate straight forwardly to number of slots If the MS asks for 7 bytes, the BS needs to give 2 slots thereby consuming 12 bytes (if 6 bytes per slot) the percentage of packet lost is also important

16 16 Fairness The time to converge to fairness is important since the fairness can be defined as short term or long term

17 17 Energy Consumption and Power Control the scheduler can calculate the suitable power to use for each MS depending upon their location (SNR, BER…) For mobile users, the power is very limited. Therefore, MS scheduler also needs to optimize the transmission power

18 18 Implementation Complexity & Scalability Since the BS has to handle many simultaneous connections and decisions have to be made within 5 ms WiMAX frame duration The algorithm should efficiently operate as the number of connections increases

19 19 Classification of Scheduler

20 20 Channel-Unaware Schedulers Intra-class Scheduling Intra-class scheduling is used to allocate the resource within the same class given the QoS requirements Inter-class Scheduling Inter-class is whether each traffic class should be considered separately, that is, have its own queue

21 21 Round Robin algorithm How to calculate average data rate or minimum reserved traffic at any given time? Another issue is what should be the duration of fairness? EX : the scheduler can allocate 100 bytes every frame for 10 frames or 1000 bytes every 10 th frame

22 22 Weighted Fair Queuing algorithm each connection has its own FIFO queue and the weight can be dynamically assigned for each queue the size of bandwidth request can be used to determine the weight of queue The ratio of a connection’s average data rate to the total average data rate can be used to determine the weight of the connection The minimum reserved rate can be used as the weight The pricing can be also used as a weight

23 23 Delay-based algorithms This set of schemes is specifically designed for real-time traffic such as UGS, ertPS and rtPS Largest Weighted Delay First (LWDF) chooses the packet with the largest delay to avoid missing its deadline Delay Threshold Priority Queuing (DTPQ) was proposed for use when both real-time and non real-time traffic are present

24 24 Priority-based algorithm (1/2) the priority order can be: UGS, ertPS, rtPS, nrtPS and BE or packets with the largest delay can be considered at the highest priority The direct negative effect of priority is that it may starve some connections of lower priority service classes

25 25 Priority-based algorithm (2/2) To mitigate starve problem, Deficit Fair Priority Queuing (DFPQ) with a counter was introduced to maintain the maximum allowable bandwidth for each service class The counter decreases according to the size of the packets. The scheduler moves to another class once the counter falls to zero

26 26 Channel-Aware Schedulers Fairness QoS guarantee system throughput maximization power optimization

27 27 Fairness Proportional Fairness Scheme (PFS) PFS uses the ratio of channel capacity W i (t) to the long-term throughput R i (t) in a given time window T i of queue i as the preference metric instead of the current achievable data rate R i (t) can be calculated by exponentially averaging the i th queue’s throughput in terms of T i

28 28 QoS Guarantee Modified Largest Weighted Delay First (M- LWDF) The algorithm explicitly uses both current channel condition and the state of the queue into account The scheme serves the queue j for which “ρ i W j (t)r j (t)” is maximum ρ i is a constant which could be different for different service classes W i (t) can be either the delay of the head of line packet or the queue length r i (t) is the channel capacity for traffic class i

29 29 System Throughput Maximization (1/2) Users with better channel conditions receive exponentially higher bandwidth that additional mechanisms are required if the total slots are less than the total minimum required slots

30 30 System Throughput Maximization (2/2) Another method is using a heuristic approach, minimize the unsatisfied demand (for UGS) R ij be the rate for MS i on channel j N ij be the number of slots allocated to MS i on subchannel j linear programming  O(n 3 m 3 N) heuristic  O(nmN)

31 31 Power Constraint (1/2) IEEE 802.16e standard defines Power Saving Class (PCS) type I, II and III to figure out the optimal sleep time with guaranteed service especially delay

32 32 Power Constraint (2/2) 60% voice packets from being transmitted during silent periods SID (Silence Insertion Descriptor)

33 33 Conclusion and future work Optimization for WiMAX scheduler is still an ongoing research topic polling mechanism backoff optimization overhead optimization Future research on scheduling Multiple Input Multiple Output multi-hops scenario handle base station selection and hand off


Download ppt "Scheduling in IEEE 802.16e Mobile WiMAX Networks-Key Issues and a Survey 報告者 : 李宗穎 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 27, NO. 2, FEBRUARY."

Similar presentations


Ads by Google